The textbook cost of capital (WACC) blends the cost of debt and the cost of equity: WACC = (E/V) × r_e + (D/V) × r_d × (1 − tax). The cost of equity, r_e, usually comes from another formula (CAPM): r_e = r_f + β × (r_m − r_f).
The trouble is that β jumps around, the extra return demanded for stocks is just a guess, and the exact-looking results (8.42%, 9.17%, and so on) are false precision. Shaky assumptions run through a precise formula still give a precisely wrong number.
So we use a simple quality-band approach instead. Start from the risk-free rate (the 10-year Treasury) and add a premium for bearing this business's risk:
• High quality (durable moat, predictable cash, low debt): +2.5 points • Above average (good franchise, modest ups and downs): +4.5 points • Average (ordinary business): +6.5 points • Below average (cyclical, leveraged, or no moat): +8.5 points
With the 10-year Treasury near 4.5%, those work out to roughly 7%, 9%, 11%, and 13%. But the rate now floats with the bond market, so the same business is not made to look cheap or dear purely by the interest-rate regime it is valued in.
The band is a judgment you make, not a number the model computes. It is the most opinionated input in any DCF, so making it visible and easy to change is more honest than hiding it inside a formula.
When testing how sensitive the answer is, we move the rate up and down by 2 percentage points around the chosen band. That is wider than the real uncertainty most of the time, but narrow enough to be useful.